Silanization integrating TiO 2 nanorods-carbon fiber for improving mechanical and wear-resisting behaviors of phenolic composite

Materials

The commercially available 3K plain woven carbon fiber (WCF) was supplied by Weihai Guangwei Group Co. Ltd., China. The Nitrile rubber-modified phenolic resin powder (Jinan Shengquan Hepworth Chemical Co., Ltd., Shandong, China) was used as matrix. Hydrochloric acid (HCl, AR, Sinopharm Chemical Reagent Co., Ltd.), isopropyl alcohol (Sinopharm Chemical Reagent Co., Ltd), tetrabutyl titanate (98% Ti(n-OBu)₄, Shanghai Lingfeng Chemical Reagent Co., Ltd., China), acetic acid, ethyl alcohol, acetone and 3-aminopropyltriethoxysilane (APS, KH550) were used as received. The deionized water was used during the whole study.

Preparation of TiO₂ nanorods-woven carbon fiber

Firstly, the commercially WCF was washed completely by acetone (named as desized-WCF). A certain volume of Ti(n-OBu)₄ was added dropwise into pre-mixed solution, consisting of deionized water, isopropyl alcohol and hydrochloric acid. Then the desized-WCF was soaked. Afterwards, the mixture was transferred into Teflon-lined stainless steel autoclave (volume ratio is 1/2) and treated at 423 K for 8 h. The as-prepared woven carbon fiber was washed thoroughly with deionized water and then dried completely (TWCF).

Bonding of 3-aminopropyltriethoxysilane

TWCF (1.0 g) was oxidized via concentrated HNO₃ (30 mL) at 353 K for 2 h, and washed with deionized water to obtain oxidized-TWCF. The oxidized-TWCF (1.0 g) was reacted with the mixture solution of ethyl alcohol (50 mL), deionized water (50 mL), acetic acid (2.0 mL) and APS (5.0 mL) at 343 K in an oven for 6 h. Then, the product was washed in excess ethyl alcohol and deionized water to remove the adherent APS on TWCF. The APS bonded TWCF was denoted as STWCF and the representative schematic illustration was shown as Figure 1. For control, the WCF bonded by APS only was also prepared, named as SWCF.

OPEN IN VIEWER

Figure 1.

Schematic illustration of APS-TiO₂ nanorods-carbon fiber and composite.

Fabrication of composites

WCF was immersed into phenolic ethanol solution fully and air dried adequately. Then, the dried preform was molded for 30 min under 443 K, 5 MPa via hot-press by vulcanizing machine. The Wt% for phenolic in composite was 30% ± 1%. For comparison, composites reinforced by desized-WCF, SWCF, TWCF and STWCF were named as C0, C1, C2 and C3, respectively. The detailed preparation description was shown in Fig. S1.

Analysis and characterizations

The woven carbon fiber was analyzed by X-ray diffraction meter using Cu Ka radiation (XRD, D8 Advance, Bruker Inc., Gemany), scanning electron microscope (SEM, S4800) and transmission electron microscopy (TEM, FEI Tecnai G2 F20S-TWIN). Functional groups on carbon fiber were characterized by Fourier transform infrared spectroscopy (FTIR, VERTE22) and X-ray photoelectron spectroscopy (XPS, Kratos Axis165 spectrometer). Besides, the wettability of WCF was tested by Contact Angle Meter (JC2000C5) and the thermal stability was characterized by thermogravimetric analyses (TGA, STA409PC). Also, the single-fiber tensile test was performed according to ASTM D3379-75. The tensile tests of composite, according to the Chinese national standard (GB/T1447-2005), were studied via universal testing machine (WDW-50H, China). The indentation experiment was conducted at room temperature with a micro indentation/scratch tester (MHT, NANOVEA, America) with setting load of 4 N and approaching speed of 15 µm/min. The frictional behavior for composite was performed through CFT-I multi-functional material surface performance tester, followed by the SEM and XPS analysis for worn surface.

Surface characterizations of woven carbon fiber

Figure 2 showed the surface morphologies of different woven carbon fibers. Compared to the smooth surface of desized-WCF (Figure 2(a)), TWCF (Figure 2(b)) was covered by needle-like nanorods, with the length ranged from 0.6 to 0.8 μm. In the case of STWCF (Figure 2(c)), the uniform needle-like nanorods on carbon fiber were still observed. According to the XRD patterns for woven carbon fibers, the nanorods on carbon fiber were confirmed to be the anatase phase TiO₂ (Fig. S2). Moreover, the TEM image for STWCF (Figure 2(d)) provided direct evidence for the in-situ growth of TiO₂ nanorods onto carbon fiber. According to the EDS analysis for the surface of STWCF show that, the C, Ti, Si and N elements were uniformly distributed on STWCF, indicated the homogeneous presence of APS on carbon fiber and TiO₂ nanorods.

OPEN IN VIEWER

Figure 2.

SEM images for desized-WCF (a), TWCF (b) and STWCF (c), HRTEM image for STWCF (d), and EDS results of STWCF (e–h).

Figure 3 displayed the FTIR spectra of different WCFs. The FTIR spectrum proved the surface chemical inertness of desized-WCF. For TWCF, the FTIR spectrum exhibited characteristic absorption bands at about 400 cm⁻¹ and 3440 cm⁻¹, originated from the stretching vibration of Ti–O–Ti and hydroxyl radical (–OH) on TiO₂ nanorods. Also, strong peaks associated with C=C (1641 cm⁻¹) and C–C (1373 cm⁻¹) stretching vibrations of carbon fibers were observed, which might promote the oxidization on carbon fiber. As the oxidization of TWCF, the broad band at 3440 cm⁻¹ would be assigned to the hydroxyl group stretching vibration, and two significant characteristic bonds appeared at 1340 and 1730 cm⁻¹ were attributed to the C–O and C=O stretching vibration of carboxylic groups, which facilitated the bond of APS on carbon fibers. For STWCF, the peak for C = O in carboxylic groups flattened, and new bands centered at 1265 cm⁻¹, 1129 cm⁻¹ and 732 cm⁻¹, corresponded to the Si–O–C, Si–O–Si and O–Si–O stretching vibration in siloxane. Moreover, the presence of a new band at about 1550 cm⁻¹ was assigned to the N–H bending vibration of amide. The above results verified the successful introduction of APS onto carbon fiber.

OPEN IN VIEWER

Figure 3.

FTIR spectra for woven carbon fibers.

As can be seen in the XPS survey spectra for woven carbon fiber (Figure 4(a)), two new peaks corresponding to N1s and Si2p were observed in spectrum for STWCF. Since the bond of APS on TWCF, remarkable shifts of Ti2p to lower binding energy (Figure 4(b)) compared to that for TWCF were observed, implying the strong interactions between TiO₂ nanorods and APS. ²⁷

OPEN IN VIEWER

Figure 4.

The XPS survey spectra for woven carbon fiber (a), Ti2p region for TWCF and STWCF (b), XPS signal of STWCF for C1s (c), O1s (d), Si2p (e) and N1s (f).

Further, the high resolution spectra of C1s, O1s, N1s and Si2p were analyzed systematically. As shown in Figure 4(c), the C1s spectrum was resolved into five component peaks. The peaks at about 284.75 and 288.55 eV were assigned to the C–C and C = O on carbon fiber, and the peaks at about 284.09, 285.65 and 286.72 eV were signified to the C–Si, C–N and C–O–Si.^4,27 In the case of O1s (Figure 4(d)), the common peaks at 531.96 eV (C = O) and 530.42 eV (C–O–Ti) ²⁸ were consistent with previous studies. In addition, two peaks originated from O–Si–O (531.1 eV) and Si–O–C (532.91 eV) appeared. Moreover, three binding energy peaks at 101.75, 102.23 and 102.71 eV were found from the high resolution spectrum of Si2p (Figure 4(e)), which were attributed to the bond –Si–O–C–, Si–OH and –Si–O–Si. It is worth noting that the N1s spectrum (Figure 4(f)) is decomposed into two peak at 399.83 and 401.55 eV, which represented to the N–H band of –NH₂ and –CON–, ²⁹ respectively. The results above confirmed the formation of effective covalent bonding (–CON– among the –COOH on oxidized-TWCF and the –NH₂ in APS.

As results, the improved wettability of STWCF was found (Fig. S3), ensuring the interfacial adhesion among carbon fibers, TiO₂ nanorods and matrix. Also, the bonded APS on carbon fibers prevents the thermal fatigue decomposition of carbon fiber (Fig. S4) from thermal shock occurs during sliding friction, hence enhancing the wear resistance of composite.

Mechanical properties of composite

The results of tensile strength for composite were depicted in Figure 5(a). Obviously, the tensile stress for C0 increased slowly with displacement, and the ultimate strength for C0 was only 97 MPa. In contrast, the ultimate strength for C1, C2 and C3 improved observably, without visible damage on carbon fiber (Fig. S5). In the case of C1, the ultimate tensile stress increased by 47.2% in comparison with that for C0. Meanwhile, steep slope of stress–strain curve was observed, which attributed to the strong chemical bonding among bonded APS on carbon fiber and matrix. ³⁰ When composite was reinforced by TWCF (C2), the tensile stress reached to 230 MPa. The improved tensile behavior for C2 was mainly due to the strong mechanical interlocking among TiO₂ nanorods and matrix though the slippage of TiO₂ nanorods took place during tensile process (zigzag peaks before final fracture). Clearly, STWCF significantly promoted the ultimate stress of composite, enlarged at 158.7% in comparison with that of desized-WCF. Notably, the slope of stress–strain curve for C3 was equivalent to that for C1 and visible zigzag peaks disappeared, suggested the largely improved bonding among carbon fiber, TiO₂ nanorods and matrix.

OPEN IN VIEWER

Figure 5.

The tensile test (a) and (b) indentation test results of composites and the SEM micrographs of fracture surface for (c) C0, (d) C1, (e) C2 and (f–h) C3, respectively.

Further, the indentation hardness (H_IT) and indentation modulus (E_IT) for the matrix around carbon fibers were shown in Figure 5(b), the H_IT for C0 was only 0.24 GPa, which increased to 0.39 GPa, 0.32 GPa and 0.42 GPa for C1, C2 and C3, respectively. Meanwhile, C0 had the lowest E_IT (2.08 GPa) compared to that for C1, C2 and C3. Observably, the SWCF embodied outstanding synergistic effect and C3 obtained a 148.5% increase in E_IT, reached to about 5.17 GPa. In addition, the strong and homogeneous interphase (Figure 6) for C3 was believed to be helpful for transferring external stress from the matrix to carbon fiber, thus improve the mechanical properties of composite.

OPEN IN VIEWER

Figure 6.

Schematic illustration of tensile failure for composites.

Moreover, Figure 5(c) to (h) presented the fracture surface for composites after tensile tests. In Figure 5(c), large voids and obvious fiber pull-out phenomenon were appeared, and the fiber surface was clean and the cross-section was flat, which proved the obvious interfacial failure for C0. On the contrary, the fiber/matrix debonding was nearly invisible in the fracture of C1 (Figure 5(d)), which ascribed to the increase of the interaction among matrix and APS on carbon fiber. In the case of C2 (Figure 5(e)), the bulk failure occurred though the ultimate tensile stress improved markedly. Notably, few resin remained on carbon fibers along with random TiO₂ nanorods, implied the deficient bonding among TiO₂ nanorods, carbon fiber and matrix.

Since C3 was reinforced by STWCF, a significant improvement had been observed (Figure 5(f) to (h)). Plenty of resin fragments adhered on carbon fiber (Figure 5(f)), together with rough cross-section of broken carbon fiber (Figure 5(g)). Moreover, the linear scanning EDS analysis across the fiber/matrix interface revealed the obvious interfacial reaction among bonded APS and matrix, and a transition layer was formed. In addition, it can be seen from Figure 5(h) that TiO₂ nanorods in interfacial region played effective bridging among carbon fibers and matrix, which increased ultimate strength of composites. This is mainly due to the fact that the growth of TiO₂ nanorods on the fiber surface can significantly increase the surface area of the fiber, and the nanorods grow vertically on the surface of the fiber. The full play of its “rivet” function helps to strengthen the mechanical lock between the fiber and the resin matrix; On the other hand, the chemical grafting of APS changes the fiber/matrix interface bonding mode, and the presence of chemical bonding significantly improves the interface bonding strength of the composite material. The possible damage mechanisms for composites under tensile test were shown in Figure 6.

Frictional behavior

Figure 7(a) showed the relation between frictional coefficient and sliding time for different composites under contact load of 150 N and sliding speed of 500 t/min. It was found that C0 displayed the lowest frictional coefficient, along with distinct oscillatory phenomenon during the whole test, indicating the changeful worn surface on C0. In contrary, C1 and C2 exhibited higher frictional coefficient than C0, except the minute fluctuation during sliding process. It is noteworthy that C3 presented a steady frictional coefficient, following a short run-in period (20 min). The increased and stability frictional coefficient for C3 suggested excellent friction property under repeated sliding friction. Moreover, C3 displayed the greatest frictional stability against to changed contact loads, especially under high sliding speed (Fig. S7).

OPEN IN VIEWER

Figure 7.

The variation of frictional coefficient with test time (a) and wear rates (b) under 150 N, 500 t/min of sliding for composites, and the corresponding wear scars.

In Figure 7(b), C0 displayed the maximum wear rate (4.6 × 10⁻⁸mm⁻³/(N m)) of all composites, indicating the poor wear resistance in this experiment. When composite was reinforced by SWCF (C1) and TWCF (C2), the wear rate was decreased by about 39.1% and 32.6% compared to C0, owing to the improved interface strength of composites. When composite was reinforced by the bonded APS-TiO₂ nanorods-woven carbon fiber, C3 showed dramatic anti-wear property, the wear rate was only about 1.7 × 10⁻⁸mm⁻³/(N m), a 63.0% reduction compared to C0. The variation of wear scars for composites was basically consistent with the results of wear rate.

Worn surface

The higher magnification images for worn surfaces of composites were show in Figure 8. Lots of exposed carbon fibers, broken carbon fibers and wear debris were observed on the worn surface of C0 (Figure 8(a)). Moreover, the sharp fracture and neat broken carbon fiber were found (Figure 8(e)), which revealed the weak abrasion resistance for C0. Comparing to the worn surface of C0, C1 relieved the detachment of matrix from worn surface with some fiber/resin debonding and broken fibers (Figure 8(b)). Besides, the interfacial adhesion between fractured carbon fibers and resin reflected the efficient combination of SWCF and matrix (Figure 8(f)). In the case of C2 (Figure 8(c)), most broken carbon fibers on worn surface were still bonded together though the abrasive wear took place. In addition, the ragged fracture and adhesion resin or debris were detected on broken carbon fibers (Figure 8(g)), indicating the improved mechanical anchor among TWCF and resin. When composite was reinforced by the STWCF, most carbon fibers were still embedded in the resin (Figure 8(d)). The fiber breakage, fiber cracking, wear debris, and fiber/matrix debonding were invisible on the worn surface of C3, provided direct evidence of the improvement in the wear resistance. Moreover, the homogeneous transfer film (Figure 8(h)) on worn surface was detected clearly, which was believed to be responsible to the enhanced surface fatigue resistance of composite.

OPEN IN VIEWER

Figure 8.

SEM micrographs of worn surfaces for composites: (a, e) C0, (b, f) C1, (c, g) C2, and (d, h) C3, and the XPS spectra for worn transfer film (i, j).

Further, the transfer film mentioned above was analyzed by XPS. As can be seen in Figure 8(i), the transfer film was mainly consisted of C (89.63%), O (6.90%) and Si (3.47%) elements and the chemical bonds for Si can be divided into –Si–O–Si–, Si–OH and –Si–O–C (Figure 8(j)). In comparison with the high resolution spectrum for Si2p of STWCF, the –Si–O–Si– content of the transfer film increased to 69.38%, accompanying with a remarkable decrease of –Si–O–C (only 1.93%). It can be considered that, a part of bonded APS on worn interface transformed into Si-base transfer film through the tribochemical reaction during sliding friction, which prevented the abrasion of C3 from the mechanical peeling and thermal decomposition effectively. This is mainly due to the synergistic enhancement of interface mechanical engagement and chemical bonding, which gives the sample a strong interface binding effect. In the continuous friction process, the carbon fiber is well protected by the matrix and is not easy to peel from the resin, so that the sample maintains a relatively large friction coefficient during the friction process. The worn process for C0 and C3 was displayed in Figure 9.

OPEN IN VIEWER

Figure 9.

Schematic illustration of the worn process for C0 and C3.